Visualization of mold deformation, and monitoring system based on Mahalanobis distance

Abstract

Techniques for visualizing and managing the state of equipment have become important in recent years, and they are widely used in industry both to save labor and to transmit technology. In plastic molding, for example, they are used for quality control by attaching sensors to molds. Managing the manufacture of plastic products, which are used in a wide variety of places, has become complicated due to stringent quality requirements and larger and more complex shapes. So-called DX manufacturing, which employs digital technologies such as artificial intelligence（AI）and the Internet of Things（IoT）, can make mold manufacture more efficient by monitoring mold conditions and identifying defects. In this study, with the aim of constructing a low-cost system, we attached strain gauges to molds, acquired data during operation, and visualized the data using general-purpose software（Microsoft Excel VBA）. We also used the Mahalanobis distance（MD）to construct a display system with the normal state as the unit space, and presented several examples. As a result, we were able to visualize mold status during the molding process and display MD values under both normal and abnormal conditions, and found that we had a tool that could be usefully employed to monitor the operation of a mold.